This is a competitive renewal application for the recently completed parent R01 (AA015069: CBT Treatment for Anxiety Disorder in Comorbid Alcoholics). Up to half of patients in treatment for an alcohol use disorder (AUD) have a co-occurring anxiety disorder and these patients relapse to drinking at twice the rate of other patients. To address this, the parent R01 evaluated a CBT-based program that combined three sessions of trans-diagnostic anxiety reduction therapy with three sessions designed to de-couple the cognitive and behavioral bonds linking anxiety to alcohol use (e.g., expectancies, coping drinking motives, conditioned associations). In a randomized trial with over 300 cases, the parent R01 showed that adding this CBT to a residential community-based AUD treatment significantly improved four-month alcohol outcomes (e.g., 41% relapsed) compared to adding a stress reduction control treatment (53% relapsed) and compared to a matched, non-randomized cohort undergoing the AUD treatment alone (61% relapsed). While this confirmed the primary study hypotheses in the parent R01, it must be acknowledged that the effects were not large and that the CBT alleviated only about half of the increased relapse risk associated with co- occurring anxiety disorders. Fortunately, the results of the parent R01 also point the way to modifications that are likely to significantly increase the therapeutic effect of the CBT. Specifically, the pattern of findings in the parent R01 indicates that the de-coupling therapy, rather than the anxiety reduction therapy, caused the improved alcohol outcomes. This suggests that further emphasizing the de-coupling therapy elements would increase the CBT's overall effectiveness; however, the importance of anxiety reduction for improving alcohol outcomes in these patients remains ambiguous. For example, the three sessions devoted to anxiety reduction may have been insufficient to affect alcohol outcomes in the parent R01. Alternatively, anxiety reduction therapy may be needed to work synergistically with de-coupling therapy to improve alcohol outcomes. The renewal work would use a dismantling approach to isolate the separate and interactive effects of these therapy components at two dose levels toward the goal of increasing the effectiveness of the validated but still sub-optimally performing parent R01 version of the CBT. 350 AUD patients with an anxiety disorder would be randomized to groups receiving either: 1) six sessions of CBT for anxiety reduction (CBT- AR); 2) six sessions of CBT for anxiety-alcohol de-coupling (CBT-DC); or, 3) the original CBT with its three anxiety reduction sessions and three de-coupling sessions (CBT-O). Based on the parent R01 findings indicating that the de-coupling components are the active ingredients of the CBT therapy, we predict that the best alcohol outcomes will be obtained in the CBT-DC group (six DC sessions) followed by the CBT-O (three DC sessions) further followed by the CBT-AR (zero DC sessions). This design would also indicate synergy effects between the AR and DC CBT components if the CBT-O out-performs the other two groups.